\begin{table}[htbp]\centering
\def\sym#1{\ifmmode^{#1}\else\(^{#1}\)\fi}
\caption{The effect of negative group appeals on statement evaluation (Study II)}
\begin{tabular}{l*{3}{cc}}
\hline\hline
                &\multicolumn{2}{c}{Model 1} &\multicolumn{2}{c}{Model 2} &\multicolumn{2}{c}{Model 3} \\
\hline
Value-policy appeal&    0.062         &   (0.06)&     0.16         &   (0.13)&     0.11         &   (0.12)\\
Group-policy appeal&     0.12\sym{*}  &   (0.06)&     1.08\sym{***}&   (0.13)&     0.94\sym{***}&   (0.11)\\
Group affect    &                  &         &    -0.43\sym{***}&   (0.02)&    -0.43\sym{***}&   (0.02)\\
Value-policy appeal $\times$ Group affect&                  &         &   -0.017         &   (0.02)&   -0.016         &   (0.02)\\
Group-policy appeal $\times$ Group affect&                  &         &    -0.18\sym{***}&   (0.02)&    -0.18\sym{***}&   (0.02)\\
Party affect    &                  &         &     0.12\sym{***}&   (0.01)&                  &         \\
Value-policy appeal $\times$ Party affect&                  &         &   -0.015         &   (0.02)&                  &         \\
Group-policy appeal $\times$ Party affect&                  &         &   -0.044\sym{*}  &   (0.02)&                  &         \\
\_cons          &     5.94\sym{***}&   (0.08)&     7.73\sym{***}&   (0.13)&     8.24\sym{***}&   (0.12)\\
\hline
Observations    &    18584         &         &    18584         &         &    18584         &         \\
FEs respondents &      Yes         &         &      Yes         &         &      Yes         &         \\
FEs groups      &      Yes         &         &      Yes         &         &      Yes         &         \\
FEs parties     &      Yes         &         &      Yes         &         &      Yes         &         \\
Respondents     &     3145         &         &     3145         &         &     3145         &         \\
R\textsuperscript{2}&    0.083         &         &     0.19         &         &     0.18         &         \\
\hline\hline
\multicolumn{7}{l}{\footnotesize Note: Clustered standard errors (by respondents) in parentheses.}\\
\multicolumn{7}{l}{\footnotesize \sym{*} \(p<0.5\), \sym{**} \(p<0.01\), \sym{***} \(p<0.001\)}\\
\end{tabular}
\end{table}
